A communicative, encouraging team player who presents ideas effectively .
Experienced in building Retail and Cyber Security Analytics solutions with an interest in enabling organizations to realize data driven solutions.
I enjoy trekking across the Himalayas and have trekked through Thorong La Pass (17,770 feets), the highest non-motorable mountain pass in the world.
DOB: 25th December 1996
1. Worked alongside Global A Team to assess customer requirements and design integration plans and solution deployment architectures
2. Develop log source integrations and application plugins for log ingestion and analysis from different API endpoints such as Gsuite Applications, Symantec EDR,Holmes Security tool etc. using Python.
3. Research on analytics use cases such as MITRE framework/Briant Kill chain implementation using Logpoint SIEM .
4. Worked with Dhaka Stock Exchange, Bangladesh on SIEM integration , understand their analytics requirements and fulfill them as per available Logpoint resources.
5. Trainer for Certified Logpoint Administrator, User and Director Trainings.
1. Oracle XBRi Loss Prevention Implementation for Talbots, US and HBC,Canada to detect
store level fraudulent behaviors / exceptions.
2. Development of cloud-based data warehousing product for Robling using Snowflake
including feasibility research on ETL/ BI reporting batch automation using Microsoft Azure.
3. Data Warehousing / Reporting for BELK Inc. US:
a. Planning of ETL framework and development of ETL scripts
b. Implementation of history conversion of BELK’s (Oracle RMS) to new REDM data model
c. Implemented Microstrategy (Web/Developer) for fulfilling over 40 Executive
Dashboard requirements.
Selected for scholarships on AI Micromasters course offered by Columbia University on Edx.
1. Worked as Campus Director to organize one of the Worlds Biggest Social Entrepreneurship Competition under Bill Clinton Initiative university wide.
2. Lead a team of 7 Core Members, 20+ Campus Ambassadors , 100+ Volunteers spread across 10 cities of Nepal.
Wrote a data scraper to extract historical data for NEPSE (not available through API’s) and integrated it
with daily batch updater to give me regular data feeds for stock market. Performing data analysis on this
data continuously to aid investment decisions via several technical/fundamental data analysis such as
Bollinger Bands, EMA crossovers, correlational pair analysis etc.
Implementation: Python and it’s libraries (talib, pandas, numpy, seaborn, plotly, sklearn)
Used the Ames Housing Dataset, Iowa to create and compare different ML models to predict the housing prices of different houses in Ames based on 79 different variables.
This solution is a work in progress and ranks among Top 71% out of over 5000 submitted solutions as of April 2021.
Implementation: Python and it’s libraries ( pandas, numpy, seaborn, plotly, sklearn)
Developed an image captioning system to provide visual assistance to visually impaired persons using Automatic Image Captioning based on Attention Model. I was tasked with the implementation of RNN architectures based on the input from CNN extracted features.
View Project
This project is an overview of simple EDA and some common ML algorithms to predict the chances of survival of each passenger in the unfortunate Titanic disaster based on a set of features such as Age, Class, Gender etc. In this project I evaluate the importance of input features, create new features and run a comparision of different algorithms such as SVC,Random Forest,Regression , Ensemble Voting, KNN etc. to find the optimal solution.
This solution ranks among top 79 % among 21902 submitted solutions as of April 2021.
The project uses reviews from users on IMDB to effectively train and test a NLP model that can distinguish whether a user feedback is positive or negative. It effectively uses Unigram and Bigram Models to calculate the predictions. Using Tfidf vectorizer to supress common yet useless words helped improve the accuracy.
View ProjectThis repo contains two of my initial AI projetcs : 2048 AI which creates an AI program to find optimal solutions for classic 2048 board game using MinMax algorithm. Another project 8 board puzzle aims at solving the puzzle using several algroithms such as BFS, DFS etc. and compares the space-time complexity.
View Project
Offered by Columbia University in partnership with Edx.
Issued : Jan 2017
Credential ID : 80d1ab8b2ceb4ccc9c63e30416508672
Offered by University of Michigan
Issued : Jul 2020
Credential ID : KA3BX6Z69BLW
Offered by University of Michigan
Issued : Jun 2020
Credential ID : 6SEHFM8SVRXD
Offered by University of Michigan
Issued : Jul 2020
Credential ID : 47KXYGGQTHE6
Offered by University of Michigan
Issued : Jul 2020
Credential ID : L6FEGX5E3ZVP
Find the corresponding links in the icons below
